Submission + - New Gens of bots that tough out injuries (eurekalert.org)
Taco Cowboy writes: Most of the bots that we have right now stop working properly they suffer damages, — The Yutu lunar explorer of China is a perfect example: It stopped functioning when problems developed in the mechanisms which deploy its solar panel
Researchers from the University of Wyoming, in collaboration with their colleagues from Pierre and Marie Curie University of France, have developed bots which can figure out how to continue to function within minutes, despite suffering injuries
The researchers got their inspiration from the amazing ability of animals to adapt to injury, There are many three-legged dogs that can catch Frisbees, for example, and if your ankle is sprained, you quickly figure out a way to walk despite the injury
"When injured, animals do not start learning from scratch," senior author Jean-Baptiste Mouret said. "Instead, they have intuitions about different ways to behave. These intuitions allow them to intelligently select a few, different behaviors to try out and, after these tests, they choose one that works in spite of the injury. We made robots that can do the same"
The breakthrough isn't based solely on the robots themselves — we've had robots capable of advanced movement for some time. What's key is a new algorithim that lets the robot figure out different gaits and motions much faster. Normally when a particular approach stops being effective, the robot tests various ways of moving until it finds one works
"If the robot has to search through the space of all possible behaviors," Clune said. "It’s going to be larger than the number of molecules on planet earth, it’s like finding one of a few needles in a field of haystacks"
The process can take hours, so Clune and his team found a more effective method: Giving the robot a simulated "childhood"
Before it is deployed, the robot uses a computer simulation of itself to create a detailed map of the space of high-performing behaviors. This map represents the robot's "intuitions" about different behaviors it can perform and their predicted value. If the robot is damaged, it uses these intuitions to guide a learning algorithm that conducts experiments to rapidly discover a compensatory behavior that works despite the damage. The new algorithm is called "Intelligent Trial and Error"
The scientists' robot has solved this by trying to mimic animals — by discovering which leg is broken and then then using trial and error to figure out the best way to continue walking
"Locomotion is a major challenge," Dr Iida said. "It's an issue of energy efficiency. Robots are unusually very inefficient compared to animals"
The bots from University of Wyoming is not the first one to mimic animals, as there are bots from other companies are also trying to mimic animals, such as Boston Dynamics, which is now owned by Google
It makes a variety of robots, including the internet sensation Big Dog, which can attain locomotion on a variety of different and difficult terrains
"Each behavior it tries is like an experiment and, if one behavior doesn't work, the robot is smart enough to rule out that entire type of behavior and try a new type," Cully continues. "For example, if walking, mostly on its hind legs, does not work well, it will next try walking mostly on its front legs. What's surprising is how quickly it can learn a new way to walk. It's amazing to watch a robot go from crippled and flailing around to efficiently limping away in about two minutes"
The same Intelligent Trial and Error algorithm allows robots to adapt to unforeseen situations, including adapting to new environments and inventing new behaviors. Jeff Clune explains that "technically, Intelligent Trial and Error involves two steps:
(1) creating the behavior-performance map, and
(2) adapting to an unforeseen situation"
The map in the first step is created with a new type of evolutionary algorithm called MAP-Elites. Evolutionary algorithms simulate Darwinian evolution by hosting "survival of the fittest" competitions in computer simulations to evolve artificially intelligent robots. The adaptation in the second step involves a "Bayesian optimization" algorithm that takes advantage of the prior knowledge provided by the map to efficiently search for a behavior that works despite the damage
"We performed experiments that show that the most important component of Intelligent Trial and Error is creating and harnessing the prior knowledge contained in the map," Clune says
This new technique will help develop more robust, effective, autonomous robots. Danesh Tarapore provides some examples
"It could enable the creation of robots that can help rescuers without requiring their continuous attention," he says. "It also makes easier the creation of personal robotic assistants that can continue to be helpful even when a part is broken"
While these engineers are focused on self-learning robots, others are developing robots and materials that can "heal themselves" when they are damaged
BAE Systems said recently that in the future, it could build drones that contained a lightweight fluid that would allow jets to heal themselves from damage sustained in flight, as well as on-board 3D printers that can make new parts, while a new plastic that can fix itself has been developed by engineers at the University of Illinois
Additional reports from
http://www.bbc.com/news/techno...
http://www.sciencedaily.com/re...
http://www.dailymail.co.uk/sci...
http://www.escapistmagazine.co...
http://cacm.acm.org/news/17732...
http://www.washingtonpost.com/...
A Youtube clip on the bots is available from https://www.youtube.com/watch?...
Researchers from the University of Wyoming, in collaboration with their colleagues from Pierre and Marie Curie University of France, have developed bots which can figure out how to continue to function within minutes, despite suffering injuries
The researchers got their inspiration from the amazing ability of animals to adapt to injury, There are many three-legged dogs that can catch Frisbees, for example, and if your ankle is sprained, you quickly figure out a way to walk despite the injury
"When injured, animals do not start learning from scratch," senior author Jean-Baptiste Mouret said. "Instead, they have intuitions about different ways to behave. These intuitions allow them to intelligently select a few, different behaviors to try out and, after these tests, they choose one that works in spite of the injury. We made robots that can do the same"
The breakthrough isn't based solely on the robots themselves — we've had robots capable of advanced movement for some time. What's key is a new algorithim that lets the robot figure out different gaits and motions much faster. Normally when a particular approach stops being effective, the robot tests various ways of moving until it finds one works
"If the robot has to search through the space of all possible behaviors," Clune said. "It’s going to be larger than the number of molecules on planet earth, it’s like finding one of a few needles in a field of haystacks"
The process can take hours, so Clune and his team found a more effective method: Giving the robot a simulated "childhood"
Before it is deployed, the robot uses a computer simulation of itself to create a detailed map of the space of high-performing behaviors. This map represents the robot's "intuitions" about different behaviors it can perform and their predicted value. If the robot is damaged, it uses these intuitions to guide a learning algorithm that conducts experiments to rapidly discover a compensatory behavior that works despite the damage. The new algorithm is called "Intelligent Trial and Error"
The scientists' robot has solved this by trying to mimic animals — by discovering which leg is broken and then then using trial and error to figure out the best way to continue walking
"Locomotion is a major challenge," Dr Iida said. "It's an issue of energy efficiency. Robots are unusually very inefficient compared to animals"
The bots from University of Wyoming is not the first one to mimic animals, as there are bots from other companies are also trying to mimic animals, such as Boston Dynamics, which is now owned by Google
It makes a variety of robots, including the internet sensation Big Dog, which can attain locomotion on a variety of different and difficult terrains
"Each behavior it tries is like an experiment and, if one behavior doesn't work, the robot is smart enough to rule out that entire type of behavior and try a new type," Cully continues. "For example, if walking, mostly on its hind legs, does not work well, it will next try walking mostly on its front legs. What's surprising is how quickly it can learn a new way to walk. It's amazing to watch a robot go from crippled and flailing around to efficiently limping away in about two minutes"
The same Intelligent Trial and Error algorithm allows robots to adapt to unforeseen situations, including adapting to new environments and inventing new behaviors. Jeff Clune explains that "technically, Intelligent Trial and Error involves two steps:
(1) creating the behavior-performance map, and
(2) adapting to an unforeseen situation"
The map in the first step is created with a new type of evolutionary algorithm called MAP-Elites. Evolutionary algorithms simulate Darwinian evolution by hosting "survival of the fittest" competitions in computer simulations to evolve artificially intelligent robots. The adaptation in the second step involves a "Bayesian optimization" algorithm that takes advantage of the prior knowledge provided by the map to efficiently search for a behavior that works despite the damage
"We performed experiments that show that the most important component of Intelligent Trial and Error is creating and harnessing the prior knowledge contained in the map," Clune says
This new technique will help develop more robust, effective, autonomous robots. Danesh Tarapore provides some examples
"It could enable the creation of robots that can help rescuers without requiring their continuous attention," he says. "It also makes easier the creation of personal robotic assistants that can continue to be helpful even when a part is broken"
While these engineers are focused on self-learning robots, others are developing robots and materials that can "heal themselves" when they are damaged
BAE Systems said recently that in the future, it could build drones that contained a lightweight fluid that would allow jets to heal themselves from damage sustained in flight, as well as on-board 3D printers that can make new parts, while a new plastic that can fix itself has been developed by engineers at the University of Illinois
Additional reports from
http://www.bbc.com/news/techno...
http://www.sciencedaily.com/re...
http://www.dailymail.co.uk/sci...
http://www.escapistmagazine.co...
http://cacm.acm.org/news/17732...
http://www.washingtonpost.com/...
A Youtube clip on the bots is available from https://www.youtube.com/watch?...
